lower trigger level,less peak found,what's wrong?

[LCHEN5154]

I am testing a NI5154 digitizer and the input signals are negetive pulses. I try to capture all pulses so I set a very small

trigger level, let's say 10mv, I am surprised to see much less waveforms captured than using a 50mv trigger level. Of course

when I try to increase the trigger level the found pulses will going down. The problem is what's wrong with small trigger levels?

I use a 1V vertical range and the digitizer is 8 bits. I also found these is a large difference when I change the trigger slop from

negetive to positive. As I understood there should no difference in the number of found pulses as the pulse will cross the trigger level any way.

Can anybody explain this? thanks a lot. 

[John_P1]

Hello,

I believe we spoke yesterday on the phone about this--I think I have an idea of what's going on but I would need more information about your signal to be sure.

The trigger accuracy of the 5154 is ±5% of the Full Scale Range (see the Specifications page). If you're on the 1V vertical range this accuracy is ±50 mV.

So, when you set 10mV, you actually will be triggering somewhere in between -40 mV and 60 mV.  When you set a 50 mV trigger level, you will actually be triggering somewhere in between 0V and 100 mV.  The point at which you are triggering should be fairly consistent within the specified accuracy range of the trigger.

I think you are getting something like this:

Since you mentioned your input signal is a pulse, there will be some ringing around the edges of the signal (like that shown above).  If your trigger level is close to these edges then you could trigger multiple times off of the ringing of the signal.  The 5154 has a 1 us rearm time.

If you want to trigger once per pulse, I would suggest setting a trigger level to be at the midpoint of your transition.  For example, if your pulse is from 0V to -500 mV you should set the trigger level to -250 mV (it will actually be somewhere in between -200 mV and -300 mV if you are on the 1 V range).  Alternatively, you could use Hysteresis to avoid triggering multiple times off of the same edge.

Does this explain what you are seeing?  I would have to see an image of what the signal looks like to be sure that this is what the behavior is.

Best Regards,

[LCHEN5154]

Hello John:
 I sent you a message but let me put it here again.
 I am really too sad with with the trigger accuracy. This will destroy all our measurement because
 the trigger accuracy will make our experiment not any precise any more.

 Our goal is to determine precisely the number of pulse in a short window
 and get the timing information of each detected pulse. So if the trigger
 has such a large jitter we lose control of precision of the detected pulse.

Here I attached some figures which show how the pulses I captured look
like. I also attached the vi which I used to capture the signals. This
 vi run the digitizer ni5154 to capture and save all the pulses in a cycle
model. A start trigger will start a new cycle measurement. Our goal is to
capture all pulses in a time window follow the start trigger.

 The data got from a ion source is listed in the following.

trigger_level(mv)  negative_slop_trigger positive_slop_trigger
5                      20                       8663
10                     133                      8399
15                     750                      8203
20                     2828                     7955
25                     7344                     7718
30                     8600                     7444
35                     8367                     7156
40                     8068                     7002
45                     7835                     6746
50                     7631                     6476
60                     7060                     6021
70                     6642                     5640
80                     6168                     5291
90                     5807                     4919
100                    5413                     4605


thank you for your help.

[John_P1]

Hello,

Let's keep working on this through email, this sounds very application specific and we can give better help that way.

Best Regards,